PCBUILD Archives

Personal Computer Hardware discussion List

PCBUILD@LISTSERV.ICORS.ORG

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
David Gillett <[log in to unmask]>
Reply To:
PCBUILD - Personal Computer Hardware discussion List <[log in to unmask]>
Date:
Sat, 20 Apr 2002 00:45:09 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (56 lines)
On 19 Apr 2002, at 11:11, Richard F. Bolha wrote:

> Could someone please explain the advantage of RAID? In addition to
> a UltraATA-100 or new Ultra-133? SCSI? What are the advantage for a
> home unit? A business workstation? A business file server? Articles
> regarding this subject?

  RAID (Redundant Array of Inexpensive Disks) is a family of
techniques for combining several small drives in various ways that
provide features you wouldn't get with a single larger -- and
usually, more expensive -- drive.

  The first obvious feature is "mirroring".  By sending each "write
to disk" command simultaneously to two drives, you automatically have
an up-to-date copy of everything.  If one of the drives goes bad, you
can run with just the other until you can get it replaced.  And if
the drive controller hardware takes care of sending the command/data
to both drives, you get this with essentially no additional CPU load
or performance hit.  (In a related configuration, "duplexing", the OS
needs to send each write command to two controllers (so you lose a
little performance), but you can have a controller die and still not
lose anything.)

  A second possibility is "striping".  Alternate blocks of
consecutive data are written to the drives in round-robin fashion.
It's possible to have one drive reading or writing data while another
is positioning the heads for the next operation.  You get the
illusion of a single large drive, but one that is also faster to
respond than a single drive could be.

  More complex options distribute error-correction data across the
drives.  You don't get quite the available capacity that you would
with striping, but you get more than you would with mirroring, and
you get a "drive array" where if any drive fails, things just slow
down a bit until you get it replaced and the necessary data is
regenerated -- no need to halt your applications or users.

  Traditionally, these techniques were mainly of interest to folks
running servers and using SCSI, where much of the I/O processing load
was already being done by a fairly "smart" controller.  But in recent
years, the speed of IDE/EIDE/ATAPI/UDMA type drives has continued to
climb, and their cost has continued to plummet, and a few
manufacturers have begun to introduce controllers which offer RAID
features with these drives.  ANd now these controllers have begun to
show up integrated on motherboards....

  Does the average home user need RAID?  No.  But anyone building a
high-availability server or a high-performance workstation should
maybe be considering it.

David Gillett

         PCBUILD maintains hundreds of useful files for download
                     visit our download web page at:
                  http://freepctech.com/downloads.shtml

ATOM RSS1 RSS2